计算机模式识别法在类风湿中医证侯的研究

The Study of Computer Pattern Recognition in Different Kinds of Rheumatoid Arthritis

根据计算机模式识别的马氏（Ｍａｈａｌａｎｏｂｉｓ）距离判别法，用ＢＡＳＩＣ等语言编制的ＩＢＭ－ＰＣ等微机适用的多元统计判别程序将健康人分别与类风湿性关节炎的寒热错杂型、寒湿阴络型及肝肾两虚型的发样微量元素Ｚｎ、Ｆｅ、Ｃｕ、Ｃａ及Ｚｎ／Ｃｕ五种因素进行分类判别．结果表明以Ｚｎ、Ｆｅ、Ｚｎ／Ｃｕ为特征参量，健康人与寒热错杂型类风湿性关节炎患者分类准确率为９５．６％，而分别以Ｚｎ、Ｃｕ和Ｚｎ、Ｆｅ、Ｃｕ为特征参量，健康人与寒湿阻络型和肝肾两虚型患者分类准确率分别为１００％和９０．２％，说明以Ｆｅ、Ｃｕ、Ｚｎ（或Ｚｎ／Ｃｕ）组成的微量元素谱含量，在不同中医证型的类风湿性关节炎和健康中是有差异的．

An attempt to brings forward a multivariate statistical classification decision method to determine the concentration of Zn,Fe,Cu,Ca and Zn/Cu in health controls and the patients with different kinds of Rheumatoid Arthritis(RA)is presented.The patients are diognosed according to TCM Zheng-differentiation. as cold-syndrome associated with heat-syndrome,obstruction of meridians due to cold-wetness evil.deficiency of Liver-yin and kidney-yin. Based on the mahalanobis distans dicision is made in BASIC language which is suitable for IBM-PC microcomputer.The result showed that the fight decision percentage of RA patients with Cold-syndrome associated with Heat--syndrome is 95.6%, according to the feature indices of Zn,Fe and Zn/Cu.The percentage of patients with obstruction of meridians due to cold-wetness evil and patients with deficiency of liver-yin and kidney-yin are 100% and 90.2%.according to the feature indices of Zn, Cu and Zn,Fe,Cu,The results suggested that the concentration of Fe,Cu,Zn and Zn/Cu between healthful human and three kinds of RA patients are greatly different.